Power Shift

The disaggregation of the mobile market, which began with Nokia, Ericsson and RIM challenging the entrenched position of Motorola back in the late 1990s, is shifting again. This time it’s being driven by a different kind of power play, namely physical power issues inside a device.

The biggest problem in shrinking die and pushing economies of scale in conjunction with Moore’s Law is related to power. Even with finFETs, power density increases to the point where it causes problems, and getting those issues ironed out effectively require full-system architectures and tradeoffs.

How some of the fabless players fit into this changing landscape remains to be seen, particularly in the short-term. Time-to-market pressures and cost issues historically have favored smaller, more nimble component makers over larger ones, but typically that shift occurrs when OEMs mature and fully understand where they can refine their business models to cut costs and gain a competitive edge.

In the new mobility pecking order—and this includes companies driven by wearable electronics, medical applications and automobiles as the Internet of Things takes root, as well as mobile phones and tablet computers—most of the big companies have been at this less than a decade. Some are only now dipping their toes in preparation for the IoT’s predicted explosive growth.

Samsung is probably the most established of the companies involved. Apple is a newcomer to building chips—its earliest PCs relied on Motorola chips, and more recently it has moved them over to Intel chips. The iPhone wasn’t even introduced until 2007. Google reportedly is only now developing its own chips, and there are rumblings that Amazon is working on its own chip technology.

In all of these cases, pushing to the latest process geometries with highly advanced power management schemes and software will be a challenge of mind-boggling proportions. Managing power is indeed a system problem that requires a deep understanding of use cases, software, power domains and the ability to bridge all of these ideas into a single, highly efficient device with ever-longer periods between charges. But it’s also an essential one to master in devices such as a smart watch or Google Glass, where heat is noticeable and can kill a product’s chances for success in a market.

This translates into market disruptions, and disruptions have both good and bad effects, depending on where you stand. In the short term, this will be a huge boon for tools and IP companies as the race to bring products to market quickly becomes a matter of survival. In all new markets, and especially consumer markets tied to the Internet, giants emerge overnight, and they can fall just as quickly.

In the mid-term, as these new OEMs begin to understand just how difficult this design process has become, it likely will force them into very deep partnerships with fabless chipmakers where information is swapped much more freely than it is today. Who’s left standing at the end of this shift among the fabless companies is uncertain, but some disruption is inevitable here. And in the long term, there should be opportunities for growth on every side as the market expands and companies rush to find their niche.

The growth between now and that final stage will be uneven, as most booms are, with unexpected fallout among chipmakers at every level, in every sector, and in every geography. The difference this time is that it is no longer just about the fastest, slickest or cheapest devices. It’s about the ability to manage power and heat at a reasonable cost across a wide range of devices, while still delivering cost-effective solutions to market in a timely fashion. That’s a much, much harder problem to solve.